Broad- Versus Narrow-Sample Evidence in Disclosure Regulation Studies: A Discussion of Badia, Duro, Jorgensen, and Ormazabal (2018)

Abstract

Broad-sample evidence that examines the effect of disclosure regulation is widespread in accounting research — often justified by its greater generalizability relative to narrow-sample evidence. Badia et al. (this issue) takes a different approach by focusing on the effect of disclosure rules in a narrow sample of U.S. and Canadian oil and gas companies. In this discussion, I argue that the contention that broad samples make research on disclosure regulation generalizable is a misconception and that the narrow-sample approach has many advantages and often advances knowledge more than broad-sample research. The key advantage of a narrow sample is that, everything else equal, we generally understand the institutional setting that generated the evidence better. Institutional details allow researchers to more precisely identify the changes that the regulation caused, assess the validity of untestable identification assumptions, and, as a field, it improves our ability to take a Bayesian approach to causal inferences. I illustrate these points by contrasting the evidence in Badia et al. to evidence from broad-sample studies on the effect of IFRS adoption and other major disclosure regulation.

Keywords: disclosure rules, regulation, research method

JEL Classification: M4, M48